Re: Bug#872913: libkf5declarative5: KDE refuses to be installed (libkf5declarative5 and python-pyqt5 have dependencies not satisfied)
On 24/08/2017 18:45, Lisandro Damián Nicanor Pérez Meyer wrote: Sorry if it sounds harsh, but no, you are not entitled to do so. The sole fact that you don't understand the whole transition thing and you still complain is a key indicator. I know what transition are. some are more prepared than others, that's the only point I complain. Again, another reason that confirms what I wrote above. It's called unstable for a reason, users using unstable are expected to find this issues and be able to cope with them reporting *real* bugs. If you don't expect to see this then switch to testing. When I finc bug in experimental or unstable before they it testing, on the other hand you are happy ;-) And I did find quite a few... 2) kde environment as a whole being maybe the most demanding qt test bench applications, it may had been useful to check hat kde is running fine *before* transitioning : reading FTBS bug opened, it appears that some essential kde package indeed have never been compiled with qt 5.9.1 (and current cmake but that's another story as it was done in //). Again, false. We did rebuild the key packages related to the transition. That's actually one of the reasons it took us that much time before uploading packages and asking for the transition slot. There should not be FTBS related only to qt 5.9 then or the change were not commited ... --eric
Re: Bug#872913: libkf5declarative5: KDE refuses to be installed (libkf5declarative5 and python-pyqt5 have dependencies not satisfied)
On jueves, 24 de agosto de 2017 18:29:03 -03 Eric Valette wrote: > On 24/08/2017 16:44, Lisandro Damián Nicanor Pérez Meyer wrote: > > The sole fact that you wrote this make it clear that you don't understand > > how transitions works and what they are for. Stuff was available in > > experimental but KDE stuff didn't actually needed to be compiled against > > experimental's Qt. > But some kde package were obviously because upgrading qt5.9.1 in > experimental was pulling only kde 5.28 stuff while unstable did not have > it due to dependencies while other did not recompile making kde in > experimental rather broken for months... Nothing in experimental is expected to work without issues. That's why we stage the packages there. > > And let's not forget that in the meantime we ended preparing the packages > > in experimental and got the slot for the transition both gcc and cmake > > changed bringing up new bugs. > > Granted. [snip] > As a end user (but with years and years of debian useage (>20)), I think > I'm still entitled to tell you : Sorry if it sounds harsh, but no, you are not entitled to do so. The sole fact that you don't understand the whole transition thing and you still complain is a key indicator. > 1) Potentially breaking one of the main graphical interfaces for weeks > via a dist-upgrade is painful especially when default installed kde > package manager actual settings do a dist-upgrade equivalent that > removes most kde stuff. BTW I'm not impacted but my list of hold > package keeps growing. Ok we are few months after a stable release so > that's less important because less user need unstable... Again, another reason that confirms what I wrote above. It's called unstable for a reason, users using unstable are expected to find this issues and be able to cope with them reporting *real* bugs. If you don't expect to see this then switch to testing. > 2) kde environment as a whole being maybe the most demanding qt test > bench applications, it may had been useful to check hat kde is running > fine *before* transitioning : reading FTBS bug opened, it appears that > some essential kde package indeed have never been compiled with qt 5.9.1 > (and current cmake but that's another story as it was done in //). Again, false. We did rebuild the key packages related to the transition. That's actually one of the reasons it took us that much time before uploading packages and asking for the transition slot. [snip] > > Again please refrain to make this comments in the future. > > I'm afraid I may continue because you've been too good before with very > smooths transitions because they were tested via experimental by several > people like me before landing unstable... And again: no, you are not right. We have been able to do faster transitions because not that many stuff changed in the middle. > I do appreciate a lot your work, take this critic as an hint to do even > better next time :-) I'm sorry eric, but the only thing you are generating from all this is us wasting time with this unnecesarily. If you don't like it *please* do switch to testing or refrain to complain due to this kind of issues in unstable. -- Confucius say: He who play in root, eventually kill tree. Lisandro Damián Nicanor Pérez Meyer http://perezmeyer.com.ar/ http://perezmeyer.blogspot.com/ signature.asc Description: This is a digitally signed message part.
Re: Bug#872913: libkf5declarative5: KDE refuses to be installed (libkf5declarative5 and python-pyqt5 have dependencies not satisfied)
On 24/08/2017 16:44, Lisandro Damián Nicanor Pérez Meyer wrote: The sole fact that you wrote this make it clear that you don't understand how transitions works and what they are for. Stuff was available in experimental but KDE stuff didn't actually needed to be compiled against experimental's Qt. But some kde package were obviously because upgrading qt5.9.1 in experimental was pulling only kde 5.28 stuff while unstable did not have it due to dependencies while other did not recompile making kde in experimental rather broken for months... And let's not forget that in the meantime we ended preparing the packages in experimental and got the slot for the transition both gcc and cmake changed bringing up new bugs. Granted. So please Eric, try to avoid this comments, you are both misleading users and showing that you really don't understand what's going on. As a end user (but with years and years of debian useage (>20)), I think I'm still entitled to tell you : 1) Potentially breaking one of the main graphical interfaces for weeks via a dist-upgrade is painful especially when default installed kde package manager actual settings do a dist-upgrade equivalent that removes most kde stuff. BTW I'm not impacted but my list of hold package keeps growing. Ok we are few months after a stable release so that's less important because less user need unstable... 2) kde environment as a whole being maybe the most demanding qt test bench applications, it may had been useful to check hat kde is running fine *before* transitioning : reading FTBS bug opened, it appears that some essential kde package indeed have never been compiled with qt 5.9.1 (and current cmake but that's another story as it was done in //). Seamlessly, kde 5.37 is currently landing in experimental but is also currently uninstallable for the very same root reason. Again, wrong. 5.37 stuff is in hold there due to various packages being in the NEW queue waiting for a FTP master to review them. Once they pass that queue the maintainer needs to check that everything is OK and then push it to unstable, as long as something doe snot requires a transition too. Ok. Again please refrain to make this comments in the future. I'm afraid I may continue because you've been too good before with very smooths transitions because they were tested via experimental by several people like me before landing unstable... I do appreciate a lot your work, take this critic as an hint to do even better next time :-) -- eric -- eric
Re: Bug#872913: libkf5declarative5: KDE refuses to be installed (libkf5declarative5 and python-pyqt5 have dependencies not satisfied)
On miércoles, 23 de agosto de 2017 14:55:34 -03 Eric Valette wrote: > > Various qt 5.9 packages are being uploaded to the unstable archives now, > > but as there are quite a few of them, they not all land in unstable at > > the same time. That's a consequence of running *unstable*, these things > > happen. > > > > Your best course of action (imo) is adding testing to your sources and > > downgrade the 5.9 packages to the 5.7 ones ... and probably prefer testing > > (over unstable) and maybe also stick with testing instead of unstable. > > While, I do hear that (usual excuse for breaking unstable, and telling > users to shut up), I nevertheless would like to say that qt 5.9.x has > been in experimental for over a month (beginning of July at least if > memory serves), with kde 5.28 been added later on also to experimental > and that very same dependency problem had existed identically in > experimental for as long and nobody did fix it, and then we rediscover > it in unstable with a much wider audience. I would have expected this > kind of transition being part of experimental before moving to unstable > as it has been in the past. The sole fact that you wrote this make it clear that you don't understand how transitions works and what they are for. Stuff was available in experimental but KDE stuff didn't actually needed to be compiled against experimental's Qt. And let's not forget that in the meantime we ended preparing the packages in experimental and got the slot for the transition both gcc and cmake changed bringing up new bugs. So please Eric, try to avoid this comments, you are both misleading users and showing that you really don't understand what's going on. > Seamlessly, kde 5.37 is currently landing in experimental but is also > currently uninstallable for the very same root reason. Again, wrong. 5.37 stuff is in hold there due to various packages being in the NEW queue waiting for a FTP master to review them. Once they pass that queue the maintainer needs to check that everything is OK and then push it to unstable, as long as something doe snot requires a transition too. Again please refrain to make this comments in the future. -- My favourite poem is the one that starts 'Thirty days hath September' because it actually tells you something. -- Groucho Marx Lisandro Damián Nicanor Pérez Meyer http://perezmeyer.com.ar/ http://perezmeyer.blogspot.com/ signature.asc Description: This is a digitally signed message part.
Re: Bug#872913: libkf5declarative5: KDE refuses to be installed (libkf5declarative5 and python-pyqt5 have dependencies not satisfied)
Various qt 5.9 packages are being uploaded to the unstable archives now, but as there are quite a few of them, they not all land in unstable at the same time. That's a consequence of running *unstable*, these things happen. Your best course of action (imo) is adding testing to your sources and downgrade the 5.9 packages to the 5.7 ones ... and probably prefer testing (over unstable) and maybe also stick with testing instead of unstable. While, I do hear that (usual excuse for breaking unstable, and telling users to shut up), I nevertheless would like to say that qt 5.9.x has been in experimental for over a month (beginning of July at least if memory serves), with kde 5.28 been added later on also to experimental and that very same dependency problem had existed identically in experimental for as long and nobody did fix it, and then we rediscover it in unstable with a much wider audience. I would have expected this kind of transition being part of experimental before moving to unstable as it has been in the past. Seamlessly, kde 5.37 is currently landing in experimental but is also currently uninstallable for the very same root reason. -- eric